Twenty-Fourth AAAI Conference
on Artificial Intelligence
(AAAI-10)
Workshop Program
July 11–12, 2010
Atlanta, Georgia
Sponsored by
Association for the
Advancement of Artificial Intelligence
445 Burgess Drive, Menlo Park, CA 94025-3442
650-328-3123
650-321-4457 (fax)
workshops10@aaai.org
Format
Deadlines
n
n
n
n
March 29: Submissions due
April 15: Notification of acceptance
May 4: Camera-ready copy due to organizers and AAAI
July 11–12: AAAI-10 Workshop Program
AAAI Formatting Guidelines
n
AAAI Workshop Chairs
www.aaai.org/Publications/Author/author.php
www.aaai.org
AAAI is pleased to present the AAAI-10 Workshop
program. Workshops will be held Sunday and
Monday, July 11–12, 2010 at the Westin Peachtree
Plaza in Atlanta, Georgia. Exact locations and
dates for the workshops will be determined in early spring.
The AAAI-10 workshop program includes 13
workshops covering a wide range of topics in artificial intelligence. Workshops are one day unless
noted otherwise in the individual description.
Each workshop is limited to approximately 25 to
65 participants, and participation is usually by invitation from the workshop organizers. However,
most workshops also allow general registration by
other interested individuals.
There is a separate fee for attendance at a workshop, and is discounted for AAAI-10 technical registrants. Registration information will be mailed
directly to all invited participants. (For information about AAAI-10, including registration, travel
and accommodations, see the AAAI-10 web page).
All workshop participants must preregister, and
must indicate which workshop(s) they will be attending. Workshop reports are included in the
workshop registration fee, and will be distributed
onsite during the workshop. In most cases, reports
will also be available after the conference as part
of the AAAI Press technical report series.
Submission Requirements
Submission requirements vary for each workshop,
but most key deadlines are uniform, unless otherwise noted. Submissions are due to the organizers
on March 29, 2010, except where noted. Workshop organizers will notify submitters of acceptance by April 15, 2010. Accepted camera-ready
copy is due on May 4, 2010. Please mail your submissions directly to the chair of the individual
workshop according to their directions. Do not
mail submissions to AAAI. For further information about a workshop, please contact the chair of
that workshop.
2
AAAI-10 WORKSHOPS
AAAI two-column format is often required for
workshop submissions, and is always required for
all final accepted submissions. Links to styles,
macros, and guidelines for this format are located
in the publications area of the AAAI website
Michael Beetz
Department of Informatics
Technische Universität München
beetz@cs.tum.edu
Giuseppe De Giacomo
Dipartimento di Informatica e Sistemistica
Sapienza Universita' di Roma, Italy
degiacomo@dis.uniroma1.it
Contents
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
W1: AI and Fun
W2: Bridging the Gap between Task and Motion Planning
W3: Collaboratively-Built Knowledge
Sources and Artificial Intelligence
W4: Goal-Directed Autonomy
W5: Intelligent Security
W6: Interactive Decision Theory and Game
Theory
W7: Metacognition for Robust Social Systems
W8: Model Checking and Artificial Intelligence
W9: Neural-Symbolic Learning and Reasoning
W10: PAIR: Plan, Activity, and Intent Recognition 2010
W11: StarAI — Statistical Relational AI
W12: Visual Representations and Reasoning
W13: Workshop on Abstraction, Reformulation, and Approximation
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
What are the research challenges and future
directions that should drive the community
as a whole?
What are the long-term societal impacts of
game AI and entertainment research?
What should the long-term goals of AI related to entertainment be?
How do we measure and/or model engagement, drama, aesthetics or “fun?”
What are the ways in which an intelligent
system can effect change in the human user
with respect to engagement and “fun?”
What are the applications and what are the
promising theoretical and applied techniques?
How will the ways in which we use computers or develop computer games change if we
achieve our goals?
To what extent must an intelligent system be
creative?
What is the role of AI for the designer and for
end-user generated content?
What are the linkages between entertainment
and education and how can AI support
them?
Submissions
We invite submission of position papers exploring
(4–8 pages) the topics and issues surrounding the
role, challenges, and opportunities of AI in creating fun, engaging, and entertaining experiences.
Position papers will be eligible for long or short
presentation. We also invite technical research
contributions (4–8 pages) for short presentation
and technology demonstrations (2 pages) that report or show recent advances in AI techniques that
support fun. See supplementary workshop website for submission instructions.
Organizing Committee
Mark Riedl, chair (riedl@cc.gatech.edu, Georgia
Tech), Vadim Bulitko (bulitko@gmail.com, University of Alberta), Charles Isbell (isbell@cc.gatech.edu, Georgia Tech), Ashwin Ram (ashwin@cc.
gatech.edu, Georgia Tech)
Additional Information
For additional information, please visit the supplemental workshop site (research.cc.gatech.edu/
aifun).
Artificial Intelligence and Fun
Interactive entertainment (aka computer games)
has become a dominant force in the entertainment sector of the global economy. The question
that needs to be explored in depth: what is the
role of artificial intelligence in the entertainment
sector? If we accept the premise that artificial intelligence has a role in facilitating the entertainment and engagement of humans, then we are left
with new questions:
The workshop aims to bring together a wide spectrum of researchers working on and thinking
about the role of AI in creating engaging, entertaining, “fun” experiences. Additionally, the workshop will provide for discussions between participants on significant challenges of the emerging
field. The workshop format will consist of presentations of position papers and technical research
contribution, discussion sessions, and invited
talks. A technology demonstration session will
provide a venue to show off existing and emerging
AI for fun systems.
AAAI-10 WORKSHOPS
3
Bridging the Gap between Task and Motion Planning
4
AAAI-10 WORKSHOPS
It has been a longstanding goal of AI and robotics
to build autonomous vehicles that can move
around on land, in the sea, and in the air interacting with the physical world to achieve their goals.
In recent years, the increasing availability of capable mobile platforms, manipulators, and highprecision sensors, coupled with advances in perception, localization and planning algorithms
have brought us much closer to achieving this
goal.
Robotic platforms have demonstrated autonomous navigation in large complex spaces for
prolonged periods of time while robotic manipulators have demonstrated autonomous manipulation of objects in cluttered spaces. However, effective, task-oriented motion inevitably requires a
principled approach to integrating task planning
and motion planning that is capable of operating
in real-time in dynamic and complex environments. Historically, general but discrete task planning has been considered extensively in the AI
community while specialized continuous motion
planning has been the focus in robotics.
The goal of this workshop is to investigate principled approaches to bridging the gap between
these two levels of planning, to foster the exchange of ideas between the two communities of
researchers, and to work towards developing common benchmarks and an infrastructure for evaluating the approaches to this problem.
With this goal in mind, we solicit contributions
on the topics that include (but are not limited to)
the following:
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
Combining kinematic and dynamic constraints with reasoning about tasks, time, and
resources.
Integration of discrete and continuous problem representations.
Hierarchical/multi-level planning architectures.
Incremental techniques for online planning.
Techniques for integrating task and manipulation planning.
Planning for compliant motion and motion
primitives.
Planning for cooperative manipulation with
multiple effectors.
Whole body control.
Interfaces between motion and task planners
Discussion of good benchmark problems
and an infrastructure for evaluating approaches.
Submissions
The submissions may be up to eight pages in
length, and should follow AAAI formatting guidelines. We also welcome shorter (as short as twopage) submissions such as position papers and
papers discussing/proposing common benchmarks and an infrastructure for evaluating the approaches to solving the problems that combine
task and motion planning. In particular, we encourage relevant submissions from the AAAI challenge track.
Organizing Committee
Maxim Likhachev (University of Pennsylvania),
Bhaskara Marthi (Willow Garage), Conor McGann (Willow Garage), David Smith (NASA)
AI and NLP have the potential to both exploit and
dig deeper in the mines of collective knowledge,
and to help build them, by providing tools for
helping generate more, better and consistent content. As with the previous events, we believe work
in this area should be encouraged, followed and
popularized, to promote the synergy between
repositories of user-contributed knowledge and
research in artificial intelligence.
The workshop is intended to be highly interdisciplinary. We encourage participation of researchers
from different perspectives, including (but not
limited to) machine learning, computational linguistics, information retrieval, information extraction, question answering, knowledge representation, human computer interaction and others. We
also encourage participation of researchers from
other areas who might benefit from the use of
large bodies of machine-readable knowledge.
Topics
Topics covered by this workshop include, but are
not limited to the following:
¶
¶
¶
¶
¶
¶
¶
¶
Using user-contributed knowledge as a
source of training data for AI tasks (both supervised and unsupervised)
Automatic methods for improving the quality of user contributions
Modeling tasks for human computation
Integrating different resources (for example
Wikipedia and WN/Cyc/other ontologies)
Extracting annotated data from user contributions
Enriching user contributions with new types
of structural information
User-contributed knowledge and the Semantic Web/Web 2.0
Automatic extraction and use of cross-lingual
information
¶
¶
¶
¶
¶
Computerized use of satellite Wiki projects
such as Wiktionary, Wikibooks or Wikispecies
Human computation like Amazon Mechanical Turk to help AI tasks
Data mining on collaboratively-contributed
resources
Innovative graph algorithms exploiting collaborative resources
Word Sense Disambiguation with Wikipedia,
Wiktionary, and so on.
Submissions
The review process is not double-blinded. Submissions should be regular full papers (up to 6
pages), short papers reporting on late-breaking results (up to 3 pages), and descriptions of system
demonstrations (up to 1 page). Please refer to the
AAAI author instruction page for the templates.
Organizing Committee
Vivi Nastase (EML Research), Roberto Navigli
(University of Rome “La Sapienza”), Fei Wu (University of Washington)
Additional Information
For additional information, please visit the supplemental workshop site (hal.di.uniroma1.it/
WikiAI-10).
Collaboratively-Built Knowledge Sources and AI
In recent years, collaborative endeavors facilitated
by the Internet seem to have the answer for the
knowledge acquisition bottleneck. More and
more resources and collaborative endeavors have
started to be incorporated and exploited as knowledge repositories for various tasks. Wikipedia with
its many facets and knowledge bearing structuring, the Tags associated with images in Flickr, and
question-answer collections in Yahoo! Answers
are a few examples of such information sources.
Amazon”s Mechanical Turk gives researchers access to “human computation” power, and is being
used more and more as a solution to the difficult
problems of large scale evaluations and data annotation, both crucial for the continuous development of the AI and NLP fields.
AAAI-10 WORKSHOPS
5
Goal-Directed Autonomy
How should an agent or group of collaborating
agents, designed to exhibit intelligent behavior, respond to unanticipated failures and opportunities
during plan execution in a complex (for example,
partially observable, stochastic, dynamic, continuous, multiagent) environment? We argue that they
should reason about their goals. In particular, they
should (1) detect situations that may trigger goal
reasoning, (2) explain why the situation demands
attention, (3) decide how to respond (for example, via goal(s) formulation), and (4) manage the
current set goals, which may involve tasks such as
goal interruption, transformation, resumption,
and/or deletion. This workshop will assess the
benefits and limitations of alternative GDA conceptual models, representations, and reasoning
methods, along with their evaluation and (potential) applications.
Agents are typically told what goals to pursue and
cannot modify them. Some methods (for example, for contingency planning, dynamic replanning) can respond to execution failures, but usually ignore opportunities and do not reason about
the goals themselves. GDA relaxes some common
assumptions of classical planning (for example,
static environments, fixed goals, no unpredictable
exogenous events), and requires attention to new
issues (for example, when, how, and what new
goals should be formulated?). Many application
contexts exist (for example, analysis of social/cultural behaviors, workflow processing, real-time
video games, military missions, disaster management).
Topics
¶
¶
¶
¶
¶
¶
¶
¶
6
AAAI-10 WORKSHOPS
Theoretical models
Representation and (meta-)reasoning methods
Roles for machine learning and/or other capabilities
GDA enhanced planning models
Demonstrations of utility
Evaluation/analyses
Comparing and contrasting GDA with other
approaches
Applications (and potential applications)
Format
This one-day workshop will include a comprehensive introductory presentation, example GDA scenarios, paper presentations, and breakout sessions
for groups to identify potential GDA models and
applicable methods, and their summary presentations. We will also include a panel on challenges
for designing, implementing, and evaluating GDA
systems. Additional time will be reserved for
demos, Q/A, and discussion of workshop topics/presentations. Interested and curious researchers are most welcome!
Attendance
This workshop is limited to 75 invited attendees.
Please notify coorganizer David Aha (david.aha@
nrl.navy.mil) if you wish to attend.
Submission Requirements
Please e-mail AAAI-styled PDF submissions (max
6 pages), GDA system demos (4), or letters of interest (1-2) to David Aha.
Organizing Committee
David Aha (NRL), Matthew Klenk (NRC/NRL),
Hector Munoz-Avila (Lehigh University), Ashwin
Ram (Georgia Institute of Technology), Daniel
Shapiro (ISLE)
Additional Information
For additional information, please visit the supplemental workshop site (home.earthlink.net/
~dwaha/research/meetings/aaai10-gda).
The use of techniques drawn from AI is increasingly relevant as the scale of the problem increases, in terms of the size and complexity of the networks being protected, the variety of applications
and services provided using that infrastructure,
and the sophistication of the attacks being made.
With this workshop, we hope to encourage dialogue and collaboration between the AI and Security communities. Further, we hope that this will
foster a continuing interaction. Previous workshops in this area include the ICAPS-09 workshop
on Intelligent Security, as well as two workshops
held in conjunction with the ACM Conference on
Computer and Communications Security (CCS),
in 2008 and 2009.
Topics
Topics of interest for this workshop include, but
are not limited to, the following:
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
Knowledge representation and engineering
for cyber security
Secure web services
Development of trusted software
Data mining and forensics
Automated vulnerability analysis
Automated exploit and attack generation
Automated alerting and response
Diagnosis and plan recognition
Automating security analyses and audits
Artificial immune systems
Privacy and confidentiality
Intelligent user interfaces for security applications
Security and organizational structure
Format
The emphasis for this one-day workshop will be
on discussion and interaction among the workshop participants, grounded in and motivated by
a limited number of presentations drawn from
full paper submissions, as well as an invited
speaker.
Submissions
We will accept either full papers for presentation,
formatted in AAAI style, or position papers with a
maximum length of 2 pages, outlining the author”s background, interests, and suggested topics
for or contributions to the workshop. All submissions should be e-mailed to Mark Boddy, at the email address given below, preferably in PDF or
PostScript format.
Intelligent Security
Our increasingly networked world continues to
provide new opportunities for security breaches
that have severe consequences at the personal level (identity theft, and resulting financial losses),
for businesses (theft of intellectual property, or
business plans, or costly responses to the theft of
customer data), and for governments. Computing
and the Internet have become crucial parts of the
infrastructure of almost every significant commercial or governmental enterprise. Turning off the
computers or disconnecting from the network has
become tantamount to turning off the power.
Organizing Committee
Mark Boddy (cochair)
Adventium Labs
111 Third Avenue South, Suite 100
Minneapolis, MN 55401, USA
Phone: 651-442-4109
e-mail: mark.boddy@adventiumlabs.org
Stefan Edelkamp (cochair)
Technologie-Zentrum Informatik
Am Fallturm 1, Raum 2.62
28357 Universität Bremen, Germany
Phone: +49-421-218-4676
e-mail: edelkamp@tzi.de
Robert P. Goldman (cochair)
Smart Information Flow Technologies LLC d/b/a
SIFT, LLC
211 N. First St., Suite 300
Minneapolis, MN 55401, USA
Phone: 612-384-3454
e-mail: rpgoldman@sift.info
Additional Information
For additional information, please visit the supplemental workshop site (www.tzi.de/~edelkamp
/secart).
AAAI-10 WORKSHOPS
7
Interactive Decision Theory and Game Theory
8
AAAI-10 WORKSHOPS
Decision and game theories are powerful tools
with which to design autonomous agents, and to
understand interactions in systems composed of
many such agents. Decision theory provides a general paradigm for designing agents that can operate in uncertain environments. Decision-theoretic
models use mathematical formalism to define the
properties of the agent”s environment, the agent”s
sensory capabilities, the ways the agent”s actions
change the state of the environment, and the
agent”s goals and preferences.
Game theory adds to the decision-theoretic framework the idea of multiple agents interacting within a common environment. It provides ways to
specify how agents can change the environment
and how the resulting changes impact their individual preferences. Building on the assumption
that agents are rational and self-interested, game
theory uses the notion of Nash equilibrium to design mechanisms for various forms of interaction
and communication that result in the overall system behaving in a stable, efficient, and fair manner.
Recent research has sought to merge advances in
decision and game theories to build agents that
may operate in uncertain environments shared
with other agents. This research has investigated
the adequacy of Nash equilibrium as a solution
concept, focused on epistemological advances in
game theory and expressive ways to model agents,
and looked into new solution concepts all with
the aim of designing autonomous agents that may
robustly interact with other, sophisticated agents
in both cooperative and noncooperative settings.
Topics
Topics include theoretical developments in decision theory or game theory applied to interactive
settings, theoretical developments in interactive
epistemology, representation and revision of interactive beliefs, integrating decision theory and
game theory, modeling strategic agent behavior,
behavioral game theory and evolutionary game
theory, learning in multiagent settings, rational
communication among agents, descriptions of
multiagent systems employing decision theory or
game theory, empirical evaluations of multiagent
systems employing decision theory or game theory, nonstandard variants of decision theory, position statements, and descriptions of deployed systems.
Submissions
Submit papers electronically in postscript or in
PDF to piotr@cs.uic.edu, by Monday, March 29,
2010.
Workshop Chair
Piotr Gmytrasiewicz (University of Illinois at
Chicago)
Organizing Committee
Piotr Gmytrasiewicz (primary contact)
Computer Science, University of Illinois at Chicago
Chicago, IL 60607
Phone: 312 355-1320
E-mail: piotr@cs.uic.edu
Prashant Doshi
Computer Science, University of Georgia
Athens, GA 30602
Phone: 706-583-0827
Email: pdoshi@cs.uga.edu
Karl Tuyls
Knowledge Engineering, Maastricht University
The Netherlands
Email: k.tuyls_NOSPAM_@_NOSPAM_maastrichtuniversity.nl
Program Committee
Brahim Chaib-draa (Laval University), Daniel Kudenko (University of York), David Pynadath (ISI),
Ed Durfee (University of Michigan), Frank Thujisman (Maastricht University), Kate Larson (University of Waterloo), Maathijs Spaan (Institute Superior Tecnico, Lisbon), Marius Silaghi (Florida
Institute of Technology), Peter McBurney (University of Liverpool), Praveen Parchuri (Carnegie
Mellon University), Sandip Sen (University of Tulsa), Simon Parsons (Brooklyn College), Yaa'kov
"Kobi" Gal (Ben Gurion University), Yifeng Zeng
(Aalborg University), Ekhlas Sonu, (publications
chair and webmaster) (University of Georgia).
Additional Information
For additional information, please visit the supplemental workshop site (www.cs.uga.edu/
~sonu/IDTGT/organization.html).
Submissions
The significance of this workshop is to continue to
bring together researchers from different areas
such as multiagent systems, planning/scheduling,
case-based reasoning and cognitive science. It will
foster discussions about ongoing research in
metacognition, identify best practices and establish directions for future research and collaborations.
Anita Raja, cochair (University of North Carolina
at Charlotte) and Darsana Josyula, cochair (Bowie
State University)
Topics
Specific topics of interest include the following:
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
Bounded rationality in social systems
Centralized versus distributed metalevel control
Computational models of self and consciousness
Conflict resolution in social systems
Emotional models for social systems
Evaluation of metareasoning systems
Human metacognition and metamemory
Integration of metalevel control and monitoring
Learning agents and metareasoning
Logical introspection and reflective logic programming
Metacognitive architectures for social systems
Metaexplanation and self-explanation
Metalevel control in social systems
Metalevel monitoring in social systems
Multiagent coordinated metareasoning
Representing metacognitive social laws
Role of state abstraction in metareasoning
Self-adaptive systems and autonomic computing
Theoretical models of metareasoning
The submissions should not exceed 8 pages in the
AAAI style, either in PostScript or PDF format.
Submissions must be e-mailed to both chairs (anraja@uncc.edu and darsana@cs.umd.edu) by the
deadline period. Short position statements are also accepted.
Organizing Committee
Program Committee
Michael L. Anderson (Franklin and Marshall College), Guido Boella (University of Torino),
Michael Cox (DARPA), Ashok Goel (Georgia
Tech), Sarit Kraus (Bar-Ilan University), Bob
Laddaga (BBN), Victor Lesser (University of Massachusetts), Tim Oates (University of Maryland,
Baltimore County), Don Perlis (University of
Maryland, College Park), Matthew Schmill (University of Maryland, Baltimore County), Shlomo
Zilberstein (University of Massachusetts)
Additional Information
For additional information, please visit the supplemental workshop site (www.cs.umd.edu/
~darsana/AAAI10MRSS).
Metacognition for Robust Social Systems
The focus of this workshop is on the design considerations, issues and challenges in using
metacognition to improve the robustness of social
systems that include purely artificial entities or
both humans and software agents. Metacognition
is the process of thinking about thinking itself. It
is composed of both metalevel control of cognitive activities and the introspective monitoring of
such activities to evaluate and to explain them.
Format
This two-day workshop will include a number of
short paper presentations, thematically organized
discussion sessions, a breakout problem-solving
session with discussion, and two speakers. We will
also include panel discussions after each group of
paper presentations so that the audience can ask
follow up questions that compare and contrast
the related positions.
AAAI-10 WORKSHOPS
9
Model Checking and Artificial Intelligence
Model checking is the process of determining
whether a logic formula is satisfied by a model.
For many logics of interest, model checking can be
efficiently automated. This has led to widespread
interest in model checking as a technique for verifying properties of systems, and the development
of model checking tools (for example, SMV, Uppaal, PHAVer, and SPIN).
The success of model checking in the computer
aided verification community has led to a growth
of interest in the use of model checking in artificial intelligence. Automated verification technologies are increasingly relevant for safety and reliability of autonomous systems. There has been a
strong interest in this area from, for example
NASA, which has applied it in the context of the
Mars rovers and other autonomous robotics systems. On the other hand, model checking, in particular if viewed in the wider context of system verification, falsification and development, has
recently benefited from the use of AI techniques,
for example search heuristics, abstraction techniques, and constraint satisfaction (particularly
SAT solving, which underlies “bounded model
checking”). One of the principal benefits of model checkers in verification is their ability to return
error traces when the specification is false. Dually,
such traces can be viewed as plans for falsifying
the specification: this duality means that there is a
close relationship with planning. In directed model checking, AI planning techniques are applied in
the search for error traces.
The MoChArt workshop brings together researchers from AI and model checking. Apart from
presentations of accepted papers, the program will
include an invited talk. We expect around 25 participants. Previous editions of the workshop were
held in Patras in 2008 (as a satellite workshop of
ECAI), Riva del Garda in 2006 (ECAI), San Francisco in 2005 (CONCUR), Acapulco in 2003 (IJCAI), and Lyon in 2002 (ECAI).
Topics
Topics of interest include (a more detailed list can
be found on the workshop webpage):
¶
¶
¶
¶
¶
¶
10
AAAI-10 WORKSHOPS
Application of model checking techniques to
AI problems.
Model checking and AI logics.
Relations between different techniques used
in the two fields for similar purposes (for example, reducing state explosion).
New model checking techniques specifically
for AI problems.
Exploitation of AI techniques in model
checking.
Software tools for model checking in AI.
¶
Model checking for verification of AI systems.
Submissions
Preliminary papers and papers on applications are
strongly encouraged. Note that this workshop will
accept submissions until March 31, and will issue
notifications to authors by April 23, 1010. All other deadlines are the same. Submissions must be
no more than 15 pages in length. Papers must be
submitted through the EasyChair web-based conference management system: follow the link from
the workshop web page. All papers will be peer reviewed.
Chairs
Ron van der Meyden
School of Computer Science and Engineering
University of New South Wales
meyden@cse.unsw.edu.au
Jan-Georg Smaus
Institut für Informatik
Albert-Ludwigs-Universität Freiburg
smaus@informatik.uni-freiburg.de
Program Committee
Rajeev Alur (University of Pennsylvania), Massimo Benerecetti (Universitá Napoli “Federico II”),
Alessandro Cimatti (IRST, Trento), Stefan
Edelkamp
(Universität
Bremen),
Enrico
Giunchiglia (Universitá Genova), Patrice Godefroid (Microsoft Research, Redmond), Aarti Gupta
(NEC Laboratories America, Princeton), Klaus
Havelund (NASA Jet Propulsion Laboratory and
Caltech), Orna Kupferman (Hebrew University),
Marta Kwiatkowska (University of Oxford),
Alessio Lomuscio (Imperial College London),
Charles Pecheur (Université Catholique de Louvain), Doron Peled (Bar Ilan University), Jussi
Rintanen (NICTA and Australian National University), Michael Wooldridge (University of Liverpool)
Additional Information
For additional information, please visit the supplemental workshop website (mochart.informatik.uni-freiburg.de).
Neural-symbolic systems combine the statistical
nature of learning and the logical nature of reasoning. The Workshop on Neural-Symbolic Learning and Reasoning is intended to create an atmosphere of exchange of ideas, providing a forum for
the presentation and discussion of the key topics
related to neural-symbolic integration.
Topics
¶
¶
¶
¶
¶
¶
¶
¶
The representation of symbolic knowledge
by connectionist systems
Learning in neural-symbolic systems
Extraction of symbolic knowledge from
trained neural networks
Reasoning in neural-symbolic systems
Biological inspiration for neural-symbolic
integration
Integration of logic and probabilities in
neural networks
Structured learning and relational learning
in neural networks
Applications in robotics, simulation, fraud
prevention, semantic web, fault diagnosis,
bioinformatics, and so on.
Submission
Researchers and practitioners are invited to submit original papers that have not been submitted
for review or published elsewhere. Submitted papers must be written in English and should not exceed 6 pages in the case of research and experience
papers, and 3 pages in the case of position papers
(including figures, bibliography and appendices)
in AAAI-10 format as described in the AAAI-10 author instructions. Submissions are not anonymous. All submitted papers will be judged based
on their quality, relevance, originality, significance, and soundness. Papers must be submitted
via easychair in PDF format at www.easychair.org/
conferences/?conf=nesy10.
Presentation
Accepted papers will have to be presented during
the workshop. The workshop will include extra
time for audience discussion of the presentation
allowing the group to have a better understanding
of the issues, challenges and ideas being presented.
Organizing Committee
Artur d'Avila Garcez (City University London,
UK), Pascal Hitzler (Wright State University, USA),
Luis Lamb (Universidade Federal do Rio Grande
do Sul, Brazil)
Additional Information
For additional information, please visit the supplemental workshop site (www.neural-symbolic.org/NeSy10).
Neural-Symbolic Learning and Reasoning
Artificial intelligence researchers continue to face
huge challenges in their quest to develop truly intelligent systems. The recent developments in the
field of neural-symbolic computation bring an opportunity to integrate well-founded symbolic artificial intelligence with robust neural computing
machinery to help tackle some of these challenges.
AAAI-10 WORKSHOPS
11
PAIR: Plan, Activity, and Intent Recognition
Plan recognition, activity recognition, and intent
recognition all involve making inferences about
other actors from observations of their behavior,
that is, their interaction with the environment and
with each other. The observed actors may be software agents, robots, or humans. This synergistic
area of research combines and unifies techniques
from user modeling, machine vision, intelligent
user interfaces, human/computer interaction, autonomous and multiagent systems, natural language understanding, and machine learning. It
plays a crucial role in a wide variety of applications including assistive technology, software assistants, computer and network security, behavior
recognition, coordination in robots and software
agents, and e-commerce and collaborative filtering.
This diversity of applications and disciplines,
while producing a wealth of ideas and results, has
contributed to fragmentation in the field, as researchers publish relevant results in a wide spectrum of journals and conferences.
This workshop seeks to bring together researchers
and practitioners from diverse backgrounds, to
share in ideas and recent results. It aims to identify important research directions and opportunities for synthesis and unification. Contributions
are sought in the following areas of research:
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
12
AAAI-10 WORKSHOPS
Plan, activity, intent, or behavior recognition
Adversarial planning, opponent modeling
Modeling multiple agents, modeling teams
User modeling on the web and in intelligent
user interfaces
Acquaintance models
Plan recognition and user modeling in marketplaces and e-commerce
Intelligent tutoring systems (ITS)
Machine learning for plan recognition and
user modeling
Personal software assistants
Social network learning and analysis
Monitoring agent conversations (overhearing)
Observation-based coordination and collaboration (teamwork)
Multiagent plan recognition
Observation-based failure detection
Monitoring multiagent interactions
Uncertainty reasoning for plan recognition
Commercial applications of user modeling
and plan recognition
Representations for agent modeling
Modeling social interactions
Inferring emotional states
Reverse engineering and program recognition
Programming by demonstration
Imitation
Due to the diversity of disciplines engaging in this
area, related contributions in other fields are also
welcome.
Submissions
We welcome submissions describing either relevant work or proposals for discussion topics that
will be of interest to the workshop. Submissions
are accepted in PDF format only, using the AAAI
formatting guidelines. Submissions must be no
longer than eight pages in length, including references and figures. Please e-mail submissions to gitars@eecs.ucf.edu.
Organizing Committee
Christopher Geib (primary contact) (University of
Edinburgh, cgeib@inf.ed.ac.uk), David Pynadath
(USC/ISI, pynadath@isi.edu), Hung Bui (SRI,
bui@ai.sri.com), Gita Sukthankar (University of
Central Florida, gitars@eecs.ucf.edu)
Additional Information
For additional information, please visit the supplemental workshop site (people.ict.usc.edu/~pynadath/PAIR-2010).
Recent years have seen an explosion of successes
in combining probability and (subsets of) first-order logic respectively programming languages and
databases in several subfields of AI: reasoning,
learning, knowledge representation, planning,
databases, NLP, robotics, vision, and so on. Nowadays, we can learn probabilistic relational models
automatically from millions of inter-related objects. We can generate optimal plans and learn to
act optimally in uncertain environments involving millions of objects and relations among them.
Exploiting shared factors can speed up messagepassing algorithms for relational inference but also for classical propositional inference such as
solving SAT problems. We can even perform lifted
probabilistic inference avoiding explicit state enumeration by manipulating first-order state representations directly.
So far, however, the researchers combining logic
and probability in each of these subfields have
been working mostly independently. We believe
the current situation actually provides us with an
opportunity for attempts at synthesis, forming a
common core of problems and ideas, and crosspollinating across subareas. We would like to explore the minimal perturbations required for each
of the AI subfields to start using statistical relational (SR) techniques.
Thus, the goal of the StarAI workshop is to reach
out to the general field of AI and to explore what
might be called statistical relational AI. We seek to
invite researchers in all subfields of AI to attend
the workshop and to explore together how to
reach the goals imagined by the early AI pioneers.
Submissions
We anticipate a one-day workshop with about 40
participants, position statements and technical
papers, two invited speakers, and a panel discussion. Those interested in attending should submit
either a technical paper (AAAI style, 6 pages maximum) or a position statement (AAAI style, 2
pages maximum) in PDF format via the CMT site
linked from the supplemental workshop homepage. All submitted papers will be carefully peerreviewed by multiple reviewers and low-quality or
off-topic papers will not be accepted.
Organizing Committee
Kristian Kersting (University of Bonn, Fraunhofer
IAIS, Germany, kristian.kersting@iais.fraunhofer.de), Stuart Russell (University of California
Berkeley, USA, russell@cs.berkeley.edu), Leslie
Pack Kaelbling (MIT, USA, lpk@csail.mit.edu),
Alon Halevy (Google, USA, halevy@google.com),
Sriraam Natarajan (University of Wisconsin at
Madison, USA, natarasr@biostat.wisc.edu),
Lilyana Mihalkova (University of Maryland at College Park, USA, lily@cs.umd.edu)
The program committee consists of over 40 illustrious researchers from different subfields of AI including but not limited to machine learning, natural language programming, planning, ILP, SRL,
knowledge representation, vision, robotics, databases, and bioinformatics.
Additional Information
For additional information, please visit the supplemental workshop site (www.biostat.wisc.edu/
~natarasr/starai.html).
Statistical Relational Artificial Intelligence (StarAI)
Much has been achieved in the field of AI, yet
much remains to be done if we are to reach the
goals we all imagine. One of the key challenges
with moving ahead is closing the gap between logical and statistical AI. Logical AI has mainly focused on complex representations, and statistical
AI on uncertainty. Intelligent agents, however,
must be able to handle both the complexity and
the uncertainty of the real world.
AAAI-10 WORKSHOPS
13
Visual Representations and Reasoning
Visual representations and reasoning play an important role in human problem solving, modeling, and design. Although the ability to think like
a human long has been a goal of AI, today’s AI
agents nonetheless are limited in their visual reasoning. Advances in this area may (1) enable
more extensive autonomous reasoning in visual
domains, (2) foster deeper computational support
for and understanding of human problem solving, modeling, and design, and (3) promote more
intense use of visual representations in humanmachine interaction. These technological goals
raise basic theoretical issues such as the precise
role of visual reasoning in intelligence and the relationship between visual reasoning and perceptual processes. Drawing participants from diverse research communities such as AI, HCI, cognitive
science, learning science and design science, this
interdisciplinary workshop aims to describe and
discuss the latest scientific research that may inform and influence progress towards these goals.
Submissions
Topics
Thomas Barkowsky (University of Bremen), Randall Davis (Massachusetts Institute of Technology), Ronald Ferguson (SAIC), Gabriela Goldschmidt (Technion – Israel Institute of
Technology), Mateja Jamnik (University of Cambridge), N. Hari Narayanan (Auburn University),
Nancy Nersessian (Georgia Institute of Technology), Patrick Olivier (Newcastle University), Steven
Tanimoto (University of Washington), Paul Thagard (University of Waterloo), Barbara Tversky
(Stanford University)
Topics of interest to this workshop include, but
are not limited to the following:
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
¶
Cognitive architectures
Comparisons of visual versus propositional
approaches
Diagrammatic reasoning
Educational applications
Formal theories of visual representation
Mental imagery in cognition
Nultimodal representations and reasoning
High-level perception
Sketch understanding
Spatial representations and reasoning
Visual media
Visual representations and mental models
Visual representations in creativity and design
Visual representations in human culture
Visual similarity and analogy
Format
This workshop will be a one-day meeting with a
combination of full and short paper presentations, along with one or two keynote speakers.
Each session will be followed by a panel discussion.
Attendance
Attendance at the workshop will be limited to
about 40 participants, to encourage active discussion.
14
AAAI-10 WORKSHOPS
We welcome two types of submissions: full papers
(6-8 pages) and shorter extended abstracts or position papers (2-3 pages). Submissions must follow the standard AAAI two-column format, and
PDF versions should be submitted via EasyChair.
Please direct all other inquiries to aaai10vrr@
easychair.org
Organizing Committee
Keith McGreggor, cochair (primary contact)
Design and Intelligence Lab
School of Interactive Computing,
Georgia Institute of Technology
85 Fifth Street NW, Atlanta, GA 30332
Maithilee Kunda, cochair (mkunda@gatech.edu),
Ellen Do (ellendo@gatech.edu), Ashok Goel
(goel@cc.gatech.edu)
Program Committee
Additional Information
For additional information, please visit the supplemental workshop site (dilab.gatech.edu/AAAI10-VRR-Workshop).
Submissions
It has been recognized since the inception of AI
that abstractions, problem reformulations, and
approximations (ARA) are central to human common-sense reasoning and problem solving and to
the ability of systems to reason effectively in complex domains. ARA techniques have been used in
a variety of problem-solving settings and application domains, primarily to overcome computational intractability by decreasing the combinatorial costs associated with searching large spaces. In
addition, ARA techniques are also useful for
knowledge acquisition and explanation generation.
Gregory Provan (g.provan@cs.ucc.ie, University
College Cork, Ireland) Ashish Sabharwal (sabhar@cs.cornell.edu, Cornell University, USA)
Topics
Topics of interest include all aspects of abstraction, reformulation and approximation, including
(but not limited to) the following:
¶
¶
¶
¶
¶
New techniques for automatically constructing and selecting appropriate ARA methods
Frameworks that unify and classify ARA techniques
Empirical and theoretical studies of the costs
and benefits of ARA
Applications of ARA to search, constraint satisfaction, deterministic and probabilistic
planning, theorem proving, logic programming, game playing, parallel and distributed
search, distributed data and knowledge
bases, internet search and navigation, knowledge compilation, knowledge acquisition,
knowledge reformulation, simulation, design, diagnosis and control of physical systems (including mobile robots), automatic
programming, analogical reasoning, casebased reasoning, reasoning under uncertainty, reinforcement learning, machine learning,
and speed-up learning
Fielded applications demonstrating the benefits of ARA to a variety of real-world domains.
The workshop will consist of an invited talk, oral
presentations,
a
poster
session,
and
discussion/brainstorming sessions. Submissions
are sought both for new work in the area of ARA
as well as for work recently published or soon to
be published in another conference or journal; for
submissions of the latter kind, the authors must
clearly state the venue of publication. The workshop proceedings will appear as an AAAI Technical Report, which are citable archival proceedings.
Submissions must be in AAAI format and be 2–6
pages in length. Please submit as PDF via email to
sabhar@cs.cornell.edu.
Organizing Committee
Program Committee
Chris Beck (University of Toronto, Canada),
Vadim Bulitko (University of Alberta, Canada),
Berthe Choueiry (University of Nebraska-Lincoln,
USA), Fausto Giunchiglia (University of Trento,
Italy), Mike Genesereth (Stanford University,
USA), Robert Holte (University of Alberta, Canada), Ian Miguel (University of St Andrews, UK),
Michael Lowry (NASA, USA), Gregory Provan
(University College Cork, Ireland), Wheeler Ruml
(Palo Alto Research Center, USA), Ashish Sabharwal (Cornell University, USA), Lorenza Saitta
(Universita del Piemonte Orientale, Italy), Sven
Koenig (University of Southern California, USA),
Toby Walsh (University of New South Wales, Australia), Jean-Daniel Zucker (Universite Paris 13 /
UR 079 Geodes, France)
Additional Information
For additional information, please or visit the
supplemental workshop site (www.cs.cornell.edu/
~sabhar/workshops/wara-2010) or send an e-mail
to the organizers.
Workshop on Abstraction, Reformulation, and Approximation
The 2010 Workshop on Abstraction, Reformulation, and Approximation (WARA-2010) will be
held in conjunction with AAAI-10 with the goal of
providing a forum for intensive interaction among
researchers in all areas of Artificial Intelligence
and Computer Science with an interest in the different aspects of abstraction, reformulation, and
approximation techniques. The aim and scope of
this workshop are similar to an independent symposium called SARA. The diverse backgrounds of
participants of previous SARA symposia has led to
a rich and lively exchange of ideas, allowed the
comparison of goals, techniques, and paradigms,
and helped identify important research issues and
engineering hurdles. This workshop will continue
to do the same.
AAAI-10 WORKSHOPS
15